Power discom BSES employs ‘predictive learning’ for better demand forecast in Delhi; what it is, how it works

The peak power demand in the national capital can escalate up to 4800 MW, BSES said.

With an aim to get more accurate demand forecast, Delhi’s power discom BSES has employed ‘predictive learning’ technique, making use of a vast data spanned over 10 long years, The Indian Express reported. A few months back, the power discom company which is a joint venture of Reliance Infrastructure Limited and the Delhi government began using a third-party software for the same purpose a few months back, The Indian Express reporte citing an unidentified official.

The peak power demand in the national capital can escalate up to 4800 MW, the company said. It had peaked at 4511 MW in the last year. The respective peak demands were 4168 and 4125 MW In 2016 and 2015.

Also read: Share market LIVE updates: Sensex off day’s high, Nifty below 10,650; Bharti Airtel, RIL surge up to

“The process basically involves analysing the large amount of data with us relating to variables like weather, humidity, temperature, seasons, holidays, working days, storms, winds to get a more accurate prediction model. All these factors can impact power demand. We are using data going back over 10 to 12 years,” The Indian Express reported citing an unidentified official.

A company spokesperson said that the  said the technique is being employed by the power discom is somewhat like the one used by the Google Assistant tool.

“In case of Google Assistant, the predictions become more accurate with usage. In this case, the model will throw up better forecast by studying and analysing more data fed into it,” The Indian Express reported citing the company spokesperson.